Integrated platinum carbon nanotube based electrocatalyst for high efficiency proton exchange membrane fuel cells 5

It was found that after the 192 h ADT test, a total loss of 26.1% in electrochemical surface area was observed for the Pt/CNT catalyst, compared to 49.8% of the Pt/VXC72R reference catal

Chapter 5 Electrochemical Stability of Integrated Pt/CNT-based

Electrocatalyst

5.1 Introduction

Long-term stability of PEMFC electrocatalysts has been recently recognized as one of the major barriers to the commercialization of PEMFCs [1] The commercial viability of PEMFCs requires a long operation life over 40,000 hours for PEMFC stacks in the stationary cogeneration system, and at least 5,000 hours for automotive applications [2] Therefore, performance degradation of PEMFCs and degradation mechanisms of their component materials have received extensive attention recently

in PEMFC research [1] Many PEMFC researchers have reported their studies in the failure modes of PEMFCs and the degradation causes and mechanisms of PEMFC components, such as membrane degradation, Pt dissolution and precipitation, carbon support corrosion and so forth [3-7] Recently Shao et al [8] reviewed the current understanding of the durability issues of Pt-based electrocatalysts under real or simulated PEMFC conditions In addition, the approaches to improve the long-term stability of PEMFC electrocatalysts and the experimental methods to scrutinize this durability issue are also discussed in their review

According to previous durability studies on PEMFCs, it is believed that the corrosion of carbon black support is one of the major degradation mechanisms for the long-term stability of PEMFC electrocatalysts [9] Given the relatively high temperature, high potential, oxygen abundant, and aqueous acidic environment at

place in the cathode catalyst layer, causing electrical isolation and dissolution of the

Fig 5.1 Illustration of carbon support corrosion in PEMFCs

C + 2H2O CO2 + 4H+ + 4e- E 0 = 0.207 V (5-1)

Pt catalyst particles as illustrated in Fig 5.1 The corrosion reaction of carbon

materials in aqueous acid electrolytes is generalized as Eq 5-1 [10] Generally, this

reaction is thermodynamically possible at the PEMFC operating potentials whereas it

is almost negligibly slow within that potential range to cause severe performance

degradation [1] However, this reaction can be greatly expedited by the supported Pt

catalysts and it has been reported to be a prominent degradation mode during the

start-up and shut-down processes of PEMFCs [11] Due to the presence of Pt catalysts,

but also at 0.6 V during potential cycling oxidation [12] Another observation by

Kangasniemi et al [11] indicated that progressive oxidations of carbon black were

acidic environment Therefore, CNT support with higher graphitic content has been

developed in recent years for PEMFC electrocatalysts based on the speculation that

CNTs are more electrochemically stable than carbon black [13] Compared with

carbon black, several research groups have reported that CNTs are more resistant to

electrochemical oxidation than carbon black with or without Pt on them [11-14] Thus

CNTs have been proposed as a promising alternative support material for resolving

the carbon corrosion problem in PEMFCs

To evaluate long-term durability of PEMFCs under real operating conditions, in industry the main methodologies usually involve a long-term real-time operation and post-mortem examination of individual component after operation However, these durability studies and testings are very time-consuming and also costly to be commonly employed [15] Currently, several types of so called accelerated degradation tests (ADT) have been developed to expedite the life testing of PEMFCs, including (i) thermal degradation under hot air conditions, (ii) reduced humidity, (iii) open circuit cell operation and (iv) electrochemically forced aging under simulated cell conditions [16, 17] Although these ADT tests have been extensively used to perform durability studies on conventional carbon black support, only a few publications treated the durability issues of CNTs as catalyst support for PEMFCs [11-14] In 2006, Shao et al [13] investigated the durability of CNT support by means

of an ADT test in a three-electrode half-cell setup In the ADT test, they first prepared

a Pt/CNT-based electrode via the conventional ink-spread process and then placed it vertically in a chamber filled with 0.5 M H2SO4 solution The Pt/CNT-based electrode

electrolyte that mimics the PEMFC environment To carry out the ADT test, a fixed potential of 1.2 V vs RHE was applied onto the working electrode for 192 h and the electrochemical surface area loss of the electrode was characterized by cyclic voltammetry along the ADT test It was found that after the 192 h ADT test, a total loss of 26.1% in electrochemical surface area was observed for the Pt/CNT catalyst, compared to 49.8% of the Pt/VXC72R reference catalyst This dramatic difference was mainly due to the more prominent Pt particle growth and Pt dissolution from the oxidized VXC72R support In another ADT test for CNT support by Wang et al [14],

potential of 0.9 V for different periods at 60 °C The electrochemical stability of the Pt/CNT catalyst was also evaluated by CV characterization, which revealed only 37% electrochemical surface area loss for the Pt/CNT catalyst after 168 h oxidation while almost 80% of Pt surface area was lost for the Pt/VXC72R catalyst Despite the higher electrochemical stability of CNT support claimed in previous durability studies,

it should be noted that these studies were always performed under a simulated PEMFC environment while little evidence is shown in literature that CNT support is more oxidation resistant than carbon black under real PEMFC conditions The long-term stability information of Pt/CNT-based electrocatalysts from a fuel cell under real operating conditions would be very useful for the full evaluation of these electrocatalysts in terms of their electrochemical performance

In this study, a series of in situ accelerated degradation tests (ADT) were performed on the integrated Pt/CNT-based electrode to examine the electrochemical stability of the Pt/CNT-based electrocatalyst The ADT tests were conducted in the real fuel cell test system by using an asymmetric MEA with a Pt/VXC72R-based anode and a Pt/CNT-based cathode, operating at the conditions described in Section

reference and the counter electrodes simultaneously Different ADT tests, including both static and dynamic potential oxidation processes, were carried out in this study to investigate their effectiveness for evaluating the electrochemical stability of the in situ grown CNTs In situ CV characterization was used to determine the electrochemical surface area loss after the ADT tests In order for comparison, 20 wt% Pt/VXC72R

(E-TEK) and 40 wt% Pt/VXC72R (Johnson Matthey) commercial catalysts were also investigated for their electrochemical stability via the ADT tests

5.2 Comparison of Electrochemical Stability of Pt/CNT and Pt/VXC72R-based Electrocatalysts

This section mainly concentrates on the electrochemical stability evaluation of the Pt/CNT-based electrocatalyst by means of a series of ADT tests under real fuel cell conditions The ADT tests were carried out in three different modes: CV cycling between 0.1−1.2 V vs DHE, potentiostatic oxidation at 1.5 V vs DHE, and potential cycling between 0.6 and 1.8 V vs DHE The electrochemical stability evaluation was performed based on the loss of electrochemical surface area revealed by in situ CV characterization The stability data of the Pt/CNT-based electrocatalyst are shown in the following subsections in terms of the different ADT methods

5.2.1 ADT 1 – CV Cycling Oxidation

In this ADT test, the Pt/CNT-based electrode was subjected to CV cycling

of 500 cycles were performed continuously, corresponding to a continuous oxidation process of about 3 h The in situ cyclic voltammograms of the Pt/CNT-based electrode were recorded before and after every 100 CV cycles In this test, a 40 wt% Pt/VXC72R-based electrode (Johnson Matthey) was used as a reference undergoing the same oxidation process

Figure 5.2 shows the CVs of the Pt/CNT and Pt/VXC72R-based electrodes during the CV cycling oxidation It was observed that the electrochemical surface area

0.0 0.2 0.4 0.6 0.8 1.0 1.2 -80

-60

-40

-20

0 20

40

60

80

Potential / V

Before CV cycling After 100 CV cycles After 200 CV cycles After 300 CV cycles After 400 CV cycles After 500 CV cycles

-80

-60

-40

-20

0 20

40

60

80

Potential / V

Before CV cycling After 100 CV cycles After 200 CV cycles After 300 CV cycles After 400 CV cycles After 500 CV cycles

Fig 5.2 500-cycle CV cycling oxidation of (a) Pt/CNT-based catalyst

and (b) 40 wt% Pt/VXC72R-based catalyst (Johnson Matthey)

of the Pt/CNT-based electrode showed only slight shrinking after the 3 h oxidation

loss of Pt surface area By contrast, it is noticeable that the ECSA of the Pt/VXC72R-

(b) (a)

based electrode exhibited a visible drop after the same oxidation process, which was calculated as 39.2% It suggests that the in situ grown CNTs with sputter-deposited Pt catalysts are more oxidation resistant under real fuel cell operating conditions, comparing the commercial Pt/VXC72R catalyst It also suggests that the corrosion reaction of the VXC72R support was notably accelerated by the dynamic CV cycling within 0.1−1.2 V vs DHE However, the low oxidation rate of the CNT support revealed that the oxidation potential should be elevated to further accelerate the CNT oxidation thus reducing the ADT test duration

5.2.2 ADT 2 – Potentiostatic Oxidation

The second ADT test was basically a potentiostatic oxidation process in which the Pt/CNT-based electrode experienced a constant potential of 1.5 V vs DHE for 2 h, instead of a dynamic potential cycling as in the CV cycling oxidation process In this ADT test, the CV characterization for electrode electrochemical surface area was performed before and after oxidation in a 30 min interval To compare with its electrochemical stability, a 20 wt% Pt/VXC72R-based electrode (E-TEK) was used as

a reference in this test

Figure 5.3 shows the oxidation current of the two electrodes during the oxidation process It can be clearly seen that the Pt/VXC72R-based electrode exhibited a notably higher oxidation current than that of the Pt/CNT-based electrode It is known that this oxidation current consists of four main components: oxidation of permeated hydrogen from anode, oxide formation on Pt surface, double-layer charging and carbon corrosion [18] The oxidation of permeated hydrogen and oxide formation on

Pt surface usually take pace instantly at potentials higher than 1.0 V, revealing as

0 1800 3600 5400 7200 0.00

0.02

0.04

0.06

0.08

0.10

Time / s

Pt/CNT catalyst

20 wt% Pt/VXC72R catalyst (E-TEK)

  Fig 5.3 Potentiostatic oxidation of Pt/CNT and 20 wt%

Pt/VXC72R-based electrodes at 1.5 V vs DHE

sharp current pulses at the beginning of oxidation after each CV characterization The double-layer charging also occurs rapidly and remains unvaried under potential oxidation Therefore, the oxidation current change during the oxidation process can be mainly attributed to the variation of the carbon corrosion current [19] As can be seen

in Fig 5.3, at a high oxidation potential of 1.5 V, the carbon corrosion current for the commercial Pt/VXC72R catalyst gradually increased in the first 30 min oxidation, and then it slowly decreased in the subsequent 30 min oxidation while a small oxidation current was maintained in the following oxidation process This current behavior implies that the corrosion of the VXC72R support mostly occurred in the first hour of the oxidation process By contrast, the oxidation current of the CNT support remained very small while little variation was observed during the whole oxidation process It suggests that the in situ grown CNT support is more oxidation resistant than the VXC72R support under the potentiostatic oxidation at 1.5 V vs DHE

0.0 0.2 0.4 0.6 0.8 1.0 1.2 -80

-60

-40

-20

0 20 40 60 80 100

120

Potential / V

Before potentiostatic oxidation After 30min oxidation at 1.5V After 60min oxidation at 1.5V After 90min oxidation at 1.5V After 120min oxidation at 1.5V

-600

-400

-200

0 200

400

600

800

Potential / V

Before potentiostatic oxidation After 30min oxidation at 1.5V After 60min oxidation at 1.5V After 90min oxidation at 1.5V After 120min oxidation at 1.5V

  Fig 5.4 CVs of (a) Pt/CNT and (b) 20 wt% Pt/VXC72R-based

electrodes before and after potentiostatic oxidation process

Figure 5.4 presents the cyclic voltammograms of the two electrodes before and after the potentiostatic oxidation Unlike by the CV cycling oxidation ADT test, a more severe oxidation was observed for the Pt/CNT-based electrode by the 2-h potentiostatic oxidation as shown in Fig 5.4 (a), probably due to the higher oxidation potential of 1.5 V According the CVs in Fig 5.4 (a), the Pt/CNT-based electrode lost

(a)

(b)

approximately 50−60% of its original electrochemical surface area after the first 30 min oxidation process Nevertheless, further ECSA loss was not quite noticeable in the following oxidation process It suggests that the corrosion of the CNT support was rather mild after the first 30 min oxidation given that most of its vulnerable content was oxidized within the first 30 min oxidation For the Pt/VXC72R-based electrode, it can be clearly seen in Fig 5.4 (b) that its electrochemical surface area was almost completely lost after the first 30 min oxidation at 1.5 V, indicating a high oxidation rate as demonstrated in Fig 5.3 The post-oxidation CVs of the Pt/VXC72R-based electrode revealed a notably enlarged double-layer charging region, suggesting a severely damaged electrode structure where electron transfer was greatly impaired [11] Moreover, a pair of redox peaks were found in the potential range of 0.5−0.7 V

in the post-oxidation CVs of the Pt/VXC72R-based electrode, corresponding to the oxidation and reduction of the quinone (Q) and hydroquinone (HQ) species formed on the oxidized carbon material surface [11, 20, 21] The oxygen content on a carbon material surface derived from Q and HQ species can be used to evaluate the oxidation resistivity of carbon materials based on their redox peak areas [11] The larger the redox peak areas of the Q and HQ species, the more easily the carbon material can be oxidized It can be clearly seen that the Pt/VXC72R-based electrode exhibited rather pronounced Q-HQ redox peak areas by the potentiostatic oxidation whereas those of the Pt/CNT-based electrode were not visible during the oxidation process It further confirmed that the Pt/CNT-based catalyst is more oxidation resistant than the Pt/VXC72R-based catalyst under a high oxidation potential of 1.5 V in the real fuel cell environment The results also suggest that the potentiostatic oxidation test at 1.5

V is more efficient than the CV cycling oxidation test that carbon corrosion was visibly accelerated by the high oxidation potential as shown for both two electrodes